The Impact of Corporate Sustainability on Organizational Processes and Performance
Robert G. Eccles, Ioannis Ioannou, and George Serafeim
Abstract
We investigate the effect of corporate sustainability on organizational processes and performance. Using a
matched sample of 180 US companies, we find that corporations that voluntarily adopted sustainability
policies by 1993 – termed as High Sustainability companies – exhibit by 2009, distinct organizational
processes compared to a matched sample of firms that adopted almost none of these policies – termed as
Low Sustainability companies. We find that the boards of directors of these companies are more likely to
be formally responsible for sustainability and top executive compensation incentives are more likely to be
a function of sustainability metrics. Moreover, High Sustainability companies are more likely to have
established processes for stakeholder engagement, to be more long-term oriented, and to exhibit higher
measurement and disclosure of nonfinancial information. Finally, we provide evidence that High
Sustainability companies significantly outperform their counterparts over the long-term, both in terms of
stock market as well as accounting performance.

Robert G. Eccles is a Professor of Management Practice at Harvard Business School. Ioannis Ioannou is an
Assistant Professor of Strategy and Entrepreneurship at London Business School. George Serafeim is an Assistant
Professor of Business Administration at Harvard Business School, contact email: gserafeim@hbs.edu. Robert Eccles
and George Serafeim gratefully acknowledge financial support from the Division of Faculty Research and
Development of the Harvard Business School. We would like to thank Christopher Greenwald for supplying us with
the ASSET4 data. Moreover, we would like to thank Cecile Churet and Iordanis Chatziprodromou from Sustainable
Asset Management for giving us access to their proprietary data. We are grateful to Chris Allen, Jeff Cronin,
Christine Rivera, and James Zeitler for research assistance. We thank Ben Esty, David Larcker (discussant), Joshua
Margolis, Costas Markides, Jeremy Stein (discussant), Catherine Thomas, and seminar participants at Boston
College, the NBER conference on the ―Causes and Consequences of Corporate Culture‖, Cardiff University, Saint
Andrews University, International Finance Corporation, Columbia University, INSEAD and the Business and
Environment Initiative at Harvard Business School for helpful comments. We are solely responsible for any errors in
this manuscript.
1
Electroniccopy
copyavailable
available at:
at: http://ssrn.com/abstract=1964011
Electronic
http://ssrn.com/abstract=1964011
1. Introduction
Neoclassical economics and several management theories assume that the corporation’s objective is profit
maximization subject to capacity (or other) constraints. The key agent in such models is the shareholder,
acting as the ultimate residual claimant who provides the necessary financial resources for the firm’s
operations (Jensen and Meckling, 1976; Zingales, 2000). However, there is substantial variation in the
way corporations actually compete and pursue profit maximization. Different corporations place more or
less emphasis on the long-term versus the short-term (Brochet et al., 2011); care more or less about the
impact of externalities from their operations on other stakeholders (Paine, 2004); focus more or less on
the ethical grounds of their decisions (Paine, 2004); and assign relatively more or less importance on
shareholders compared to other stakeholders (Freeman et al., 2007). For example, Southwest Airlines has
identified employees and Novo Nordisk patients (i.e., their end customers) as their primary stakeholder.
During the last 20 years, a relatively small but growing number of companies have voluntarily
integrated social and environmental issues in their business models and daily operations (i.e. their
strategy) through the adoption of related corporate policies.1 Such integration of environmental and
social issues into a company’s business model raises a number of fundamental questions for scholars of
organizations. Does the governance structure of firms that adopt environmental and social policies differ
from that of other firms and, if yes, in what ways? Do such firms have distinct stakeholder engagement
processes and adopt different time horizons for their decision-making? In what ways are their
measurement and reporting systems different? Finally, what are the performance implications of
integrating social and environmental issues into a company’s strategy and operations?
Some scholars argue that companies can ―do well by doing good‖ (Godfrey, 2005; Margolis et
al., 2007; Porter and Kramer, 2011) based on the assumption that meeting the needs of other stakeholders
– e.g. employees through investment in training - directly creates value for shareholders (Freeman et al.,
2010, Porter and Kramer, 2011). It is also based on the assumption that by not meeting the needs of other
stakeholders, companies can destroy shareholder value because of consumer boycotts (e.g., Sen et al.,
2001), the inability to hire the most talented people (e.g., Greening and Turban 2000), and by paying
potentially punitive fines to governments.
On the other hand, other scholars argue that adopting
environmental and social policies can destroy shareholder wealth (e.g., Friedman 1970; Clotfelter 1985;
Navarro 1988; Galaskiewicz 1997). In its simplest form, their argument is that sustainability may simply
be a type of agency cost: managers receive private benefits from embedding environmental and social
1
During the same period many more companies were active in corporate social responsibility (CSR) as an ancillary
activity. However, many of these companies did not necessarily implement or were unable to implement CSR as a
central strategic objective of the corporation. Moreover, CSR has diffused broadly in the business world only in the
last seven years (Eccles and Krzus, 2010).
2
Electroniccopy
copyavailable
available at:
at: http://ssrn.com/abstract=1964011
Electronic
http://ssrn.com/abstract=1964011
policies in the company’ strategy, but doing so has negative financial implications for the organization
(Baloti and Hanks 1999; Brown et al., 2006). Moreover, these companies might experience a higher cost
structure (e.g. paying their employees living rather than market wages). Consequently, the argument
continues, companies that do not operate under such additional environmental and social constraints will
be more competitive and as a result, will be more successful in a highly competitive environment. In fact,
this hypothesis is well captured in Jensen (2001) who states: ―Companies that try to do so either will be
eliminated by competitors who choose not to be so civic minded, or will survive only by consuming their
economic rents in this manner.‖ (p. 16).
In this study, we shed light on the organizational and performance implications of integrating
social and environmental issues into a company’s strategy and business model through the adoption of
corporate policies. The overarching thesis of our work is that organizations that voluntarily integrate
environmental and social policies in their business model represent a fundamentally distinct type of the
modern corporation, characterized by a governance structure that in addition to financial performance,
accounts for the environmental and social impact of the company, a long-term approach towards
maximizing inter-temporal profits, an active stakeholder management process, and more developed
measurement and reporting systems. Empirically, we identify 90 companies – we term these as High
Sustainability companies - with a substantial number of environmental and social policies adopted for a
significant number of years (since the early to mid-1990s), reflecting strategic choices that are
independent and in fact, far preceded the current hype around sustainability issues (Eccles and Krzus,
2010). Subsequently, we use propensity score matching in 1993 to identify 90 comparable firms that
adopted almost none of these policies; we term these as Low Sustainability companies. In the year of
matching, the two groups operate in exactly the same sectors and exhibit statistically identical size, capital
structure, operating performance, and growth opportunities. By generating matched pairs of firms as early
as 1993, we are therefore able to not only focus on long-term organizational implications but also to
introduce a long time lag between our independent and dependent variables, thus mitigating the likelihood
of bias that could arise from reverse causality.
Consistent with our expectations, we find that the group of High Sustainability firms is
significantly more likely to assign responsibility to the board of directors for sustainability and to form a
separate board committee for sustainability. Moreover, they are more likely to make executive
compensation a function of environmental, social, and external perception (e.g., customer satisfaction)
metrics. This group is also significantly more likely to establish a formal stakeholder engagement process
where risks and opportunities are identified, the scope of the engagement is defined ex ante, managers are
trained in stakeholder engagement, key stakeholders are identified, results from the engagement process
are reported both internally and externally, and feedback from stakeholders is given to the board of
3
directors. This set of High Sustainability firms also appears to be more long-term oriented: they have an
investor base with a larger proportion of long-term oriented investors and they communicate more longterm information in their conference calls with sell-side analysts. Since information is a crucial asset that
a corporation needs to have for effective strategy execution by management, as well as the effective
monitoring of this execution by the board, we find that High Sustainability firms are more likely to
measure information related to key stakeholders such as employees, customers2, and suppliers — and to
increase the credibility of these measures by using auditing procedures. We also find that High
Sustainability firms not only measure but also disclose more nonfinancial (e.g., environmental, social, and
governance) data. Our findings suggest that to a large extent the adoption of these sustainability policies
reflects by 2009 their underlying institutionalization within and across the organization rather than
reflecting acts undertaken as part of ―greenwashing‖ and ―cheap talk‖ (Marquis and Toffel, 2011).
Importantly, we show that there is significant variation in subsequent accounting and stock
market performance across the two groups of firms in the long run. In particular, we track corporate
performance for 18 years and find that High Sustainability firms outperform Low Sustainability firms
both in stock market as well as accounting performance. Using a four-factor model to account for
potential differences in the risk profile of the two groups, we find that annual abnormal performance is
higher for the High Sustainability group compared to the Low Sustainability group by 4.8% (significant at
less than 5% level) on a value-weighted base and by 2.3% (significant at less than 10% level) on an equal
weighted-base. We find that High Sustainability firms also perform better when we consider accounting
rates of return, such as return-on-equity (ROE) and return-on-assets (ROA) and that this outperformance
is more pronounced for firms that sell products to individuals (i.e., business-to-customer [B2C]
companies), compete on the basis of brand and reputation, and make substantial use of natural resources.
Finally, using analyst forecasts of annual earnings we find that the market underestimated the future
profitability of the High Sustainability firms compared to the Low Sustainability ones.
2. Sample Selection and Summary Statistics
To understand the effects of integrating social and environmental issues in an organization’s business
model, we first need to identify companies that have explicitly placed a high level of emphasis on
employees, customers, products, the community, and the environment as part of their strategy and
business model. Moreover, we need to find firms that have adopted these policies for a significant number
of years prior to CSR becoming widespread, to reduce the possibility of potential measurement error due
to the inclusion of firms that are either ―greenwashing‖ or adopting these policies purely for public
relations and communications reasons. Finally, by identifying firms based on policy adoption decisions
that were made a sufficiently long time ago - thus introducing a long lag between our independent and
2
Although we find directionally consistent results for customers, our results are not statistically significant.
4
dependent variables - we mitigate the likelihood of biases that could potentially arise from reverse
causality.
We identify two groups of firms: those that have and those that have not adopted a
comprehensive set of corporate policies related to the environment, employees, community, products, and
customers. The complete set of these policies is provided in the Appendix. Examples of policies related to
the environment include whether the company has a policy to reduce emissions, uses environmental
criteria in selecting members of its supply chain, and whether the company seeks to improve its energy or
water efficiency. Policies related to employees include whether the company has a policy for diversity
and equal opportunity, work-life balance, health and safety improvement, and favoring internal
promotion. Policies related to community include corporate citizenship commitments, business ethics,
and human rights criteria. Policies related to products and customers include product and services quality,
product risk, and customer health and safety. The Thomson Reuters ASSET4 database, which has already
been used in the literature (Cheng, Ioannou and Serafeim, 2012; Ioannou and Serafeim, 2012), provides
data on the adoption or non-adoption of these policies, for at least one year, for 775 US companies in
fiscal years 2003 to 2005.3 We eliminate 100 financial institutions, such as banks, insurance companies,
and finance firms, because their business model is fundamentally different and many of the environmental
and social policies are not likely to be applicable or material to them. For the remaining 675 companies
we construct an equal-weighted index of all policies (Sustainability Policies) that measures the percentage
of the full set of identified policies that a firm is committed to in each year.
Moreover, we track over time the extent of adoption of these policies for those organizations that
score at the top quartile of Sustainability Policies. We do so by reading published reports, such as annual
and sustainability reports, and visiting corporate websites to understand the historical origins of the
adopted policies. Furthermore, we conducted more than 200 interviews with corporate executives to
validate the historical adoption of these policies. At the end of this process, we were able to identify 90
organizations that adopted a substantial number of these policies in the early to mid-90s. We label this set
of firms as the High Sustainability group. Of the remaining 78 firms, 70 firms adopted these policies
gradually over time mostly after 1999. For eight firms we were unable to identify the historical origins of
these policies. The High Sustainability group had adopted by the mid-90s on average 40% of the policies
identified in the Appendix, and by the late 2000s almost 50%. Subsequently, we match each of the firms
in the High Sustainability group with a firm that scores in the lowest two quartiles of Sustainability
3
Founded in 2003, ASSET4 was a privately held Swiss-based firm, acquired by Thomson Reuters in 2009. The firm
collects data and scores firms on environmental and social dimensions since 2002. Research analysts of ASSET4
collect more than 900 evaluation points per firm, where all the primary data used must be objective and publicly
available. Typical sources include stock exchange filings, annual financial and sustainability reports, nongovernmental organizations’ websites, and various news sources. Every year, a firm receives a z-score for each of
the pillars, benchmarking its performance with the rest of the firms in the database.
5
Policies. Firms in those two quartiles have, on average, adopted only 10% of the policies, even by the late
2000s. These same firms had adopted almost none of these policies in the mid-90s. Because we require
each firm in the High Sustainability group to be in existence since at least the early 1990s, we impose the
same restriction for the pool of possible control firms. After this filter, the available pool of control firms
is 269.
We implement a propensity score matching process to produce a group of control firms that looks
as similar as possible to our High Sustainability group. The match is performed in 1993 because this is the
earliest year that we can confirm any one of the firms included in the High Sustainability group had
adopted these policies. To ensure that our results are not particularly sensitive to the year we choose for
the matching procedure, we redo the matching in 1992 and 1994. In any one year less than 5% of the
matched pairs change, suggesting that the year we choose for matching does not affect our final sample
set. We match each High Sustainability firm with a control firm that is in the same industry classification
benchmark subsector (or sector if a firm in the same subsector is not available), by requiring exact
matching for the sector membership. We use as covariates in the logit regression the natural logarithm of
total assets (as a proxy for size), ROA,4 asset turnover (measured as sales over total assets), market value
of equity over book value of equity (MTB), as a proxy for growth opportunities, and leverage (measured
as total liabilities over total assets). We use propensity score matching without replacement and closest
neighbor matching.5 Size and asset turnover load with a positive and highly significant coefficient in the
logit regression (untabulated results). The coefficient on MTB is positive and weakly significant. The
coefficients on leverage and ROA are both insignificant. We label the set of control firms that are selected
through this process as the Low Sustainability group.
Table 1 Panel A, shows the sector composition of our sample and highlights that a wide range of
sectors are represented. Panel B shows the average values of several firm metrics across the two groups in
the year of matching. The High Sustainability group has on average, total assets of $8.6 billion, 7.86%
ROA, 11.17% ROE, 56% leverage, 1.02 turnover, and 3.44 MTB. Similarly, the matched firms (i.e., the
Low Sustainability group) have on average, total assets of $8.2 billion, 7.54% ROA, 10.89% ROE, 57%
leverage, 1.05 turnover, and 3.41 MTB. None of the differences in the averages across the two groups are
statistically significant, suggesting that the matching process worked effectively. The two groups are
statistically identical in terms of sector membership, size, operating performance, capital structure, and
4
We also used ROE as a measure of performance and all the results were very similar to the results reported in this
paper. We also included other variables such as stock returns over the past one, two or three years but none of them
was significant.
5
Using a caliper of 0.01 to ensure that none of the matched pairs is materially different reduces our sample by two
pairs or four firms. All our results are unchanged if we use that sample of 176 firms.
6
growth opportunities. Moreover, the two groups have very similar risk profiles: both the standard
deviation of daily returns and the equity betas are approximately equal.
3. Corporate Governance
The responsibilities of the board of directors and the incentives provided to top management are two
fundamental attributes of the corporate governance system. Boards of directors perform a monitoring and
advising role and ensure that management is making decisions in a way that is consistent with
organizational objectives. Top management compensation systems align managerial incentives with the
goals of the organization by linking executive compensation to key performance indicators that are used
for measuring corporate performance (Govindarajan and Gupta, 1985). Ittner, Larcker, and Rajan (1997)
showed that the use of nonfinancial metrics in annual bonus contracts is consistent with an
―informativeness‖ hypothesis, where nonfinancial metrics provide incremental information regarding the
manager’s action choice.
Therefore, we posit that for organizations that consider environmental and social objectives as
core issues for their strategy and operations, the board of directors is more likely to have direct
responsibility over such issues; it is also more likely that top management compensation will be a
function of sustainability metrics in addition to other traditional financial performance metrics. To test
these predictions we analyze proprietary data provided to us by Sustainable Asset Management (SAM).
SAM collects the relevant data and constructs the Dow Jones Sustainability Index. Once a year, SAM
initiates and leads an independent sustainability assessment of approximately 2,250 of the largest
corporations around the world. The SAM Corporate Sustainability Assessment is based on the annual
SAM Questionnaire, which consists of an in-depth analysis based on approximately 100 questions on
economic, environmental, and social issues, with a particular focus on companies’ potential for long-term
value creation. The questionnaire is designed to ensure objectivity by limiting qualitative answers through
predefined multiple-choice questions. In addition, companies must submit relevant information to support
the answers provided. The SAM Questionnaires are distributed to the CEOs and heads of investor
relations of all the companies in the starting universe. The completed company questionnaire, signed by a
senior company representative, is the most important source of information for the assessment.
Table 2, Panel A shows the governance data items that SAM provided to us for fiscal year 2009,
as they relate to the board of directors and the executives’ incentive systems. We find results that are
consistent with our predictions. Fifty three percent of the firms in the High Sustainability group assign
formal responsibility around sustainability to the board of directors. In contrast, only 22% of the firms in
the Low Sustainability group hold the board accountable for sustainability. Similarly, 41% (15%) of the
firms in the High Sustainability group (Low Sustainability group) form a separate board committee that
deals with sustainability issues. The responsibilities and duties of a sustainability committee include both
7
assisting the management with strategy formulation and reviewing periodically sustainability
performance. For example, the principal functions of the sustainability committee of the Ford Corporation
include assisting management in the formulation and implementation of policies, principles, and practices
to foster the sustainable growth of the company on a global basis and to respond to evolving public
sentiment and government regulation in the area of GHG emissions and fuel economy and CO 2
regulation. Other functions include assisting management in setting strategy, establishing goals, and
integrating sustainability into daily business activities, reviewing new and innovative technologies that
will permit the company to achieve sustainable growth, reviewing partnerships and relationships that
support the company’s sustainable growth, and reviewing the communication and marketing strategies
relating to sustainable growth.
Another important governance feature is the set of metrics that are linked to senior executive
compensation. The two groups differ significantly on this dimension as well: High Sustainability firms
are more likely to align senior executive incentives with environmental, social, and external (i.e.,
customer) perception performance metrics, in addition to financial metrics. Of the firms in the High
Sustainability group, 18%, 35%, and 32% link compensation to environmental, social, and external
perception metrics, respectively. In contrast, only 8%, 22%, and 11% of the firms in the Low
Sustainability group link compensation to environmental, social, and external perception metrics. Firms in
the High Sustainability group are more likely to use monetary incentives to help executives focus on
nonfinancial aspects of corporate performance that are important to the firm. For example, Intel has
linked executive compensation to environmental metrics since the mid-90s, and since 2008 Intel links all
employees’ bonuses to environmental metrics. The 2010 metrics focused on carbon emission reductions
in Intel’s operations and energy-efficiency goals for new products. While the environmental component
represents a relatively small portion of the overall employee bonus calculation, Intel believes that it helps
focus employees on the importance of achieving its environmental objectives.
Moreover, in Panel B we present results from a multivariate analysis of these governance
mechanisms. To avoid results overload, we construct a variable that summarizes all the mechanisms
discussed in Panel A by calculating the percentage of mechanisms that a firm has adopted. Because the
firms might look considerably different in terms of size, growth opportunities, and performance at 2009,
we control for these factors in our model by measuring them at the end of 2009. Consistent with the
results above, we find that firms in the High Sustainability group adopt significantly more of the
mechanisms described in Panel A: the coefficient on High Sustainability is positive and significant
(0.144, p-value=0.006). Larger firms and more profitable firms have more of these mechanisms, whereas
growth opportunities are not related to their adoption. Overall, the results suggest that firms included in
the High Sustainability group are characterized by a distinct governance structure: responsibility over
8
sustainability is more likely to be directly assigned to the board of directors and top management
compensation is also more likely to be a function of a set of performance metrics that critically includes
sustainability metrics.
4. Stakeholder Engagement
Since High Sustainability firms are characterized by a distinct corporate governance model that focuses
on a wider range of stakeholders as part of their corporate strategy and business model, we predict that
such firms are also more likely to adopt a greater range of stakeholder engagement practices. This is
because engagement is necessary for understanding these stakeholders’ needs and expectations in order to
make decisions about how best to address them (Freeman, 1984; Freeman et al., 2007). With regards to
stakeholder management, prior literature has suggested and empirically shown that it is directly linked to
superior financial performance by enabling firms to develop intangible assets in the form of strong longterm relationships, which can become sources of competitive advantage (e.g., Hillman and Keim, 2001).
In other words, superior stakeholder engagement is fundamentally based on the firm’s ability to establish
such relationships with key stakeholders over time. Similarly, it has been argued that when a corporation
is able to credibly commit to contracting with its stakeholders on the basis of mutual trust and cooperation
and a longer-term horizon – as opposed to contracting in an attempt to curb opportunistic behavior – then
the corporation ―will experience reduced agency costs, transactions costs, and costs associated with team
production‖ (Jones, 1995; Foo, 2007; Cheng et al., 2011). We argue therefore, that firms that have
embedded the elements of mutual trust and cooperation and the building of long-term relationships with
key stakeholders through the incorporation of social and environmental issues in their strategy and
business model will be better positioned to pursue these more efficient forms of contracting (Jones, 1995).
On the other hand, firms that have not integrated social and environmental issues are more likely to
contract on the basis of curbing opportunistic behavior and this will impede their ability to adopt a broad
range of stakeholder engagement practices.
To get a better understanding of the differences in the stakeholder engagement model across the
two groups of firms in our sample, we again use proprietary data obtained through SAM. Panel A of
Table 3 presents a comparison between the High and Low Sustainability firms across several data items
that relate to actions prior to, during, and after stakeholder engagement. In particular, each item in Table 3
measures the frequency of adoption of the focal practice within each of the two groups, and the last
column presents a significance test of the differences between them. As before, the data are for the fiscal
year of 2009. We find that High Sustainability firms are more likely to adopt practices of stakeholder
engagement for all three phases of the process (prior to, during, and after) compared to Low Sustainability
ones.
9
Prior to the stakeholder engagement process, High Sustainability firms are more likely to train
their local managers in stakeholder management practices (14.9% vs. 0%, Training), and to perform their
due diligence by undertaking an examination of costs, opportunities, and risks (31.1% vs. 2.7%,
Opportunities Risks Examination). Moreover, High Sustainability firms are more likely to identify issues
and stakeholders that are important for their long-term success (45.9% vs. 10.8%, Stakeholder
Identification). During the stakeholder engagement process itself, our analysis shows that High
Sustainability firms are more likely to ensure that all stakeholders raise their concerns (32.4% vs. 2.7%,
Concerns) and to develop with their stakeholders a common understanding of the issues relevant to the
underlying issue at hand (36.5% vs. 13.5%, Common Understanding). In addition, they are more likely to
mutually agree upon a grievance mechanism with the stakeholders involved (18.9% vs. 2.7%, Grievance
Mechanism) and to agree on the targets of the engagement process (16.2% vs. 0%, Targets).Moreover,
High Sustainability firms are more likely to pursue a mutual agreement on the type of engagement with
their stakeholders (36.5% vs. 8.1%, Scope Agreement).
Finally, we find that after the completion of the stakeholder engagement process, High
Sustainability firms are more likely to provide feedback from their stakeholders directly to the board or
other key departments within the corporation (32.4% vs. 5.4%, Board Feedback), and are more likely to
make the results of the engagement process available to the stakeholders involved (31.1% vs. 0%, Result
Reporting) and the broader public (20.3% vs. 0%, Public Reports). In sum, High Sustainability firms
appear to be more proactive, more transparent, and more accountable in the way they engage with their
stakeholders.
Moreover, in Panel B we present results from a multivariate analysis of these stakeholder
engagement mechanisms. Similar to Section 3, we construct a variable that summarizes all the
mechanisms discussed in Panel A by calculating the percentage of mechanisms that a firm has adopted.
Consistent with the results above, we find that firms in the High Sustainability group adopt significantly
more of the stakeholder engagement mechanisms described in Panel A. In general, the results of this
section confirm our predictions: High Sustainability firms are distinct in their stakeholder engagement
model in that, compared to the Low Sustainability firms they are more focused on understanding the
needs of their stakeholders, making investments in managing these relationships, and reporting internally
and externally on the quality of their stakeholder relationships.
5. Time Horizon
The previous section argued for a distinct stakeholder management model and provided evidence for the
adoption of a wider range of stakeholder engagement practices. In assessing the impact of stakeholder
engagement, previous literature has argued that the effective management of stakeholder relationships can
generate persistence of superior financial performance over the longer-term, or in the faster recovery of
10
poorly performing firms (Choi and Wang, 2009). This occurs because building good stakeholder relations
as part of a corporation’s strategy, takes time to materialize, is idiosyncratic to each corporation, and
depends on its history; such relationships are based on mutual respect, trust, and cooperation and take
time to develop. In other words, effective stakeholder engagement necessitates the adoption of a longerterm time horizon.
To date, the extant literature on ―short-termism‖ (e.g., Laverty, 1996) has shown that executive
compensation incentives that are based on short-term metrics may push managers towards making
decisions that deliver short-term performance at the expense of long-term value creation. Consequently, a
short-term focus on creating value may result in a failure to make the necessary strategic investments to
ensure future profitability. Importantly, such a short-term approach to decision-making often implies a
negative externality being imposed on various other key stakeholders. In other words, short-termism is
incompatible with extensive stakeholder engagement and a focus on stakeholder relationships. It is also
true then that the pathologies of short-termism are less likely to be suffered by corporations with a clear
focus and commitment to multiple stakeholders. Given the documented commitment of High
Sustainability firms to stakeholder engagement therefore, we further predict that they are more likely to
adopt such a longer-term approach, and that this approach will also be reflected in the type of investors
that are attracted to such corporations.6
In Panel A of Table 4 we empirically test whether High Sustainability firms are focused more on
a longer-term horizon in their communications with analysts and investors. A company communicates its
norms and values both internally and externally, and since a long-term time horizon is one key element of
integrating social and environmental issues into strategy, we would expect High Sustainability firms to
put greater emphasis on the long-term than the Low Sustainability ones do. Investors that are interested in
generating short-term results by selling their stock after it has (hopefully) appreciated will avoid investing
in long-term-oriented firms since these firms are willing to sacrifice such short-term results if doing so
will produce long-term gains. In contrast, investors who plan to hold a stock for a long period of time will
be attracted to firms that are optimizing financial performance over a longer time horizon and are less
interested in short-term performance fluctuations. For example, after Paul Polman became the CEO of
Unilever and announced the implementation of the Sustainable Living Plan while abolishing quarterly
earnings reports, ownership of Unilever’s stock by hedge funds dropped from 15% to 5% in three years,
6
We acknowledge that under some conditions the reverse may be true: investor behavior may be driving managerial
decision-making. However, in the case of sustainability policies, we argue that this is rather unlikely. Since
stakeholder relations take several years to build, the probability of a large enough shareholder base retaining
ownership for a sufficiently long amount of time in order to institute a radical corporate change towards
sustainability seems low. This line of argument would also require investors to themselves engage with the company
over a long period of time in such a way as to establish a culture of more long-term thinking which in turn, would
push the corporation towards better shareholder and other stakeholder engagement.
11
which led to reduced fluctuations in the company’s share price. To test our predictions, we use data from
Thomson Reuters Street Events to measure the extent to which the content of the conversations between a
focal corporation and sell-side analysts is comprised of long-term vs. short-term keywords. We construct
this measure following the methodology in Brochet, Loumioti, and Serafeim (2012), as the ratio of the
number of keywords used in conference calls that characterize time periods of more than one year over
the number of keywords that characterize time periods of less than one year. Second, we measure the time
horizon of the investor base of a corporation following Bushee (2001) and Bushee and Noe (2000), by
calculating the percentage of shares outstanding held by ―dedicated‖ vs. ―transient‖ investors. Bushee
(2001) classifies institutional investors using a factor and a cluster analysis approach. Transient investors
have high portfolio turnover and highly diversified portfolios. In contrast, dedicated investors have low
turnover and more concentrated holdings. We measure how long-term oriented the investor base of a firm
is by calculating the difference between the percentage of shares held by dedicated investors and the
percentage of shares held by transient investors.
The results presented in Table 4 are consistent with our predictions. We find that High
Sustainability firms are more likely to have conference call discussions with analysts whose content is
relatively more long-term as opposed to short-term focused (1.08 vs. 0.96, Long-term vs. Short-term
Discussion). In addition, High Sustainability firms are significantly more likely to attract dedicated rather
than transient investors (-2.29 vs. -5.31, Long-term vs. Short-term Investors). In Panel B of Table 4 we
present results from a multivariate analysis of these long-term oriented behaviors and characteristics and
we find consistent results. In sum, our findings suggest that High Sustainability firms are effective
communicators of their long-term approach: not only do they speak in those terms but in fact, they are
convincing long-term investors to invest in their stock.
6. Measurement and Disclosure
Measurement
Performance measurement is essential for management to determine how well it is executing on its
strategy and to make any necessary corrections (Kaplan and Norton, 2008). The quality, comparability,
and credibility of information are enhanced by internal and external audit procedures that verify the
accuracy of this information or the extent to which practices are being followed. Given that High
Sustainability firms place a greater emphasis on stakeholder engagement than the Low Sustainability
firms, we would expect the same to be true for particular key stakeholder groups including employees,
customers, and suppliers. In particular, we would expect the High Sustainability firms to place
significantly more emphasis on measuring and monitoring performance, auditing performance measures,
adherence to standards, and reporting on performance. Using the proprietary SAM data described in
Section 4, we are able to test for differences in the extent to which the two groups of firms measure, audit,
12
and report on their performance as it relates to these three stakeholder groups. Table 5 presents a
comparison between the High and Low Sustainability firms for Employees (Panel A), Customers (Panel
B), and Suppliers (Panel C). Similar to the results of previous sections, each of these three panels
measures the frequency of adoption of the focal practice within each of the two groups, and the last
column presents a significance test of the differences between them.
For Employees, we find significant differences on three of the four metrics. High Sustainability
firms are significantly more likely to measure execution of skill mapping and development strategy
(54.1% vs. 16.2%, HR Performance Indicators/Nonfinancial), the number of fatalities in company
facilities (77.4% vs. 26.3%, KPI Labor/EHS Fatalities Tracking), and the number of ―near misses‖ on
serious accidents in company facilities (64.5% vs. 26.3%, KPI Labor/EHS Near Miss Tracking). We find
no significant difference between the two groups for the percentage of companies that use health and
safety performance tracking to follow labor relations issues. This may be due to laws and regulations
requiring all firms to perform such measures (e.g., as required by the Occupational Health and Safety
Administration [OSHA]), thereby leveling the field and eliminating any potential differences that could
have been in place under conditions where such laws and regulations did not exist; the high percentages
for both groups indicate that this might be the case (95.2% vs. 89.5%, KPI Labor / EHS Performance
Tracking). These results therefore, reflect the greater commitment that High Sustainability firms make to
the employee stakeholder group.
Panel B focuses on Customers and shows the frequency of adoption of seven relevant practices.
Contrary to our expectations and in contrast to our findings regarding Employees, there is virtually no
difference between High Sustainability and control firms on any one of these metrics, although across all
metrics more firms in the High Sustainability group measure customer-related data. We note that across
both groups a very small percentage of firms have adopted these metrics. If anything, one could argue that
the relationship between effective engagement and the creation of shareholder value is even more direct
for Customers than it is for Employees; yet even in the High Sustainability group, very few are measuring
the quality of this relationship. We suggest that one possible reason for this could be the rather primitive
state of customer relationship management practices. Moreover, our data seem to suggest that these
results are linked to the ease with which these practices can be measured. For example, variables like Cost
of Service and Potential Lifetime Value are very difficult to measure with only 6.8% and 8.1%,
respectively, of the High Sustainability firms measuring this variable. The highest percentages for this
group are for Geographical Segmentation (18.9%), Customer Generated Revenues (18.9%), and
Historical Sales Trends (16.2%) which are relatively easier to measure.
In contrast to Customers, there are some significant differences between the two groups of firms
in terms of Suppliers. In particular, we examine the standards used to select and manage relationships
13
with Suppliers, which can determine the quality of the relationship they have with the firm. Panel C
shows the frequency of adoption of 11 related practices: six of these are strongly and significantly
different across the two groups with p-values of <0.001, and the rest are significantly different at p-values
<0.06. These standards fall into either environmental or social issues, or a combination of the two. In
terms of environmental issues, significantly more High Sustainability firms use environmental monitoring
systems in the certification/audit/verification process (50.0% vs. 18.2%, Environmental Management
Systems), environmental data availability by the supplier (12.3% vs. 0.0%, Environmental Data
Availability), the supplier’s environmental policies (17.4% vs. 0.0%, Environmental Policy), and the
supplier’s environmental production standards (45.6% vs. 25.7%, Environmental Production Standards)
in selecting and evaluating suppliers than do Low Sustainability firms. Similarly, on social issues for
selecting and evaluating suppliers, significantly more High Sustainability firms use human rights
standards such as forced labor, slave labor, and child labor (17.4% vs. 5.7%, Human Rights Standards),
labor standards/requirements (18.6% vs. 8.1%, Labor Standards), and occupational, health, and safety
standards (62.9% vs. 25.7%, OHS Standards). Finally, High Sustainability Firms make a greater use of
compliance to general standards, both international (12.3% vs. 0.0%, International Standards
Compliance) and domestic (14.9% vs. 8.1%, National Standards Compliance), in selecting and evaluating
Suppliers.
The reliability and credibility of performance measurement is enhanced when it is subject to some
form of objective, third-party audit or assurance. The purpose of an audit is to ensure that the appropriate
measurement standards have been applied and that the internal control and measurement systems
producing information according to these standards, are robust. Companies can also perform internal
audits whereby a separate department is responsible for verifying the numbers produced by another
department. With rare exceptions, an external or internal audit or assurance opinion is not required for
reported nonfinancial information on a company’s environmental and social performance. However,
given the greater importance that High Sustainability firms accord to nonfinancial metrics (e.g., linking
executive compensation to such metrics), we predict a greater use of assurance by High Sustainability
firms than Low Sustainability firms.
Panel D shows the frequency of adoption of 13 focal practices regarding the use of internal and
external audit and assurance procedures. For the most part, the results are marginally significant but the
one case where our hypothesis does get clear support is having an external third-party conduct an audit of
the company’s corporate sustainability report which reports on its environmental, social, and governance
performance (11.1% vs. 1.4%, Sustainability report external audit), with a p-value of 0.017. The only
other item that has any degree of statistical significance is when the company bases its performance
measurement on relevant external standards and programs, such as AccountAbility’s AA 1000 standard
14
and the Global Reporting Initiative’s G3 Guidelines; 16.2% of the High Sustainability firms do this, in
contrast to only 2.7% of the Low Sustainability ones.
We note that very few of the High Sustainability firms have implemented assurance practices: of
the 11 focal items in Panel D the highest percentage for the High Sustainability firms is 16.2%. There are
a number of reasons for why assurance procedures are so uncommon. Technologies for measuring and
auditing nonfinancial information are still in their infancy and remain at a relatively primitive state of
development compared to financial information (Simnett, Vantraelen, and Chua, 2009). This is not
surprising given that external reporting of such information only started about 10 years ago, has only
received a significant level of interest in the past five years, and even today only a small percentage of
companies are reporting this information. One of the most important and difficult to overcome barriers to
auditing nonfinancial information includes the lack of an agreed-upon set of measurement standards. This
in turn, makes it very difficult to create auditing standards. Another barrier is the lack of sophisticated
information technology systems for measuring nonfinancial performance, especially compared to the
sophisticated and robust systems developed for financial reporting. Three other barriers are important to
note.
First, traditional audit firms are in the early stages of developing the capabilities to audit
nonfinancial information. This, combined with the lack of standards and IT systems, creates the second
barrier, which is a concern that performing this function will increase their legal risk beyond the amount
they already face for performing financial audits. Third, firms which do have capabilities for auditing
nonfinancial information, such as engineering firms for environmental information and human resource
supply chain consultants for social information, lack the global scale and full range of capabilities that
would be required to serve a large corporation that wants a single group to perform this audit. While a
large number of boutique firms could be hired to do this, the aggregate transaction and coordination costs
would be high.
Finally, in Panel E we present results from a multivariate analysis of nonfinancial measurement
and assurance mechanisms. Similar to prior sections, we construct a variable that summarizes all the
mechanisms discussed in Panels A through D by calculating the percentage of mechanisms that a firm has
adopted within each of the stakeholder groups, and with regards to assurance. Consistent with the results
above, we find that firms in the High Sustainability group adopt significantly more of the nonfinancial
measurement practices described in Panels A-D: the coefficients on High Sustainability are positive and
significant for Employees and Suppliers (but not for Customers), and the same is true for the assurance
dimension.
Disclosure
Another important element is the extent to which a company is willing to be transparent in its external
reporting about its environmental and social impact. Reporting on such nonfinancial performance
15
measures to the board is an essential element of corporate governance so that the board can form an
opinion about whether management is executing the strategy of the organization well. Moreover, external
reporting of performance improves managerial accountability to shareholders and other stakeholders.
Therefore, we expect High Sustainability firms to be more transparent and to exhibit a better balance
between financial and nonfinancial information in their external reporting. We test this prediction in Panel
A of Table 6 based on four focal metrics. First, we use ESG Disclosure scores, calculated by both
Bloomberg and Thomson Reuters; it is a measure of how complete the company’s reporting is on a range
of environmental, social, and governance topics based on a scale of 0% to 100%. Table 6 compares the
(average) percentages of High and Low Sustainability firms. The average Bloomberg ESG Disclosure
score for High Sustainability firms is 29.90%, compared to 17.86% for the Low Sustainability ones. The
corresponding percentages for the Thomson Reuters ESG Disclosure score are 46.38% and 36.91%,
respectively. The Thomson Reuters ESG disclosure score screens fewer data points for the presence of
disclosure, and that is why firms tend to have better disclosure under this score. Both of these differences
are statistically significant across the two groups. We also compared the two groups in terms of the
percentage of firms whose sustainability reports cover their entire global activities, using Thomson
Reuters ASSET4 data. A more global report represents a higher level of transparency and accountability
than one focused only on a company’s home country. We again find a statistically significant difference:
41.1% of the High Sustainability firms have a global sustainability report compared to only 8.31% of the
Low Sustainability firms.
Using data provided by SAM, we also tested whether High Sustainability firms are more likely to
integrate environmental and social information with their financial reporting. Integration of environmental
and social information in financial reports is increasingly being advocated as a way to ensure that
corporations are held accountable for their impact on the environment and society (Eccles and Krzus,
2010) and in fact, it was recently mandated in South Africa. We find that 25.7% of the High
Sustainability firms integrate social information and 32.4% integrate environmental information. In
contrast, 5.4% of the Low Sustainability firms integrate social information and 10.8% integrate
environmental information. Moreover, we analyzed the difference in the balance between financial and
nonfinancial discussion in conference calls, using the Thomson Reuters Street Events conference call
database described in Section 4. We classified all words referring to items captured by the accounting
system and the stock market system as financial. We classified words that would typically be found in a
balanced scorecard (Kaplan and Norton, 1996), except for financial keywords, as nonfinancial.7 Then we
7
We identified 38 keywords as nonfinancial. Examples include customer, employee, supplier, risk management,
reputation, leadership, strategy, and brand. We identified 155 keywords as financial. Examples include sales,
earnings, gross margin, and cash flow.
16
constructed a ratio that measures the number of nonfinancial keywords over financial keywords. The
average ratio for the High Sustainability firms is 0.96, suggesting that on average these firms are using an
equal number of financial and nonfinancial keywords in their discussion with the investment community.
In contrast, the average ratio for the Low Sustainability firms is 0.68, suggesting that on average these
firms are discussing less frequently about nonfinancial aspects of the business such as employees,
customers, suppliers, and products.
Finally, in Panel B of Table 6 we present results from a multivariate analysis (OLS and logistic
models as appropriate) of these nonfinancial disclosure mechanisms. We use the variables from Panel A
as our dependent variables and we control for firm size, growth opportunities, and performance measured
at the end of 2009, as before. Consistent with the results above, we find that firms in the High
Sustainability group adopt significantly more of the nonfinancial mechanisms described in Panel A: the
coefficients on High Sustainability are positive and highly significant for all our specifications.
7. Corporate Performance
A question that we haven’t yet addressed in our study is whether firms in the High Sustainability group
under or outperform their counterparts in the Low Sustainability group. On the one hand, firms in the
High Sustainability group might underperform because they experience high labor costs by providing
excessive benefits to their employees, forego valuable business opportunities that do not fit their values
and norms (such as selling products with adverse environmental consequences), and denying to pay
bribes to gain business in corrupt countries where bribe payments are the norm. In other words, High
Sustainability companies face tighter constraints in how they can behave. Since firms are trying to
maximize profits subject to capacity constraints, tightening those constraints further can lead to lower
profitability.
On the other hand, firms in the High Sustainability group might outperform the control firms
because they are able to attract better human capital, establish more reliable supply chains, avoid conflicts
and costly controversies with nearby communities (i.e., maintain their license to operate), and engage in
more product and process innovations in order to be competitive under the constraints that the integration
of social and environmental issues places on the organization. For example, Philips has translated its
environmental commitments to product innovation around energy-efficient light bulbs and developing
solar-power lighting in sub-Saharan Africa. Similarly, as of 2010, Siemens had over €20 billion in
revenues coming from its environmental portfolio.
Empirical examinations of the link between sustainability and corporate financial performance
have resulted in contradictory findings, ranging from a positive to a negative to a U-shaped, or even to an
inverse-U shaped relation (Margolis and Walsh, 2003). According to McWilliams and Siegel (2001),
conflicting results are due to ―several important theoretical and empirical limitations‖ (p.603) of prior
17
studies; others have argued that prior work suffered from ―stakeholder mismatching‖ (Wood and Jones,
1995), the neglect of ―contingency factors‖ (e.g. Ullmann, 1985), ―measurement errors‖ (e.g. Waddock
and Graves, 1997) and omitted variable bias (Aupperle et al., 1985; Cochran and Wood, 1984; Ullman,
1985). Importantly, none of these studies has measured financial performance over long periods of time to
allow for superior sustainability performance to impact either positively or negatively on financial
performance.
To delve into the performance implications of integrating social and environmental issues into a
company’s strategy and business model we track the stock market performance of firms in both groups
from 1993 to 2010. The use of stock returns addresses concerns over reverse causality in the absence of
private information. In the presence of private information, reverse causality is a concern. For example, if
managers with private information that their firms are going to outperform in the future adopt
environmental and social policies today, then the expectation of higher stock returns is causing the
adoption of these policies. However, we believe that this explanation is unlikely for a number of reasons.
First, we are not aware of a theory suggesting that managers expecting to outperform market expectations
in the future would be more likely to adopt environmental and social policies today. More importantly,
empirical evidence suggests that managers are unable to forecast returns past 100 days (Jenter, Lewellen,
and Warner, 2011). Therefore, accurately forecasting returns over the next 3, 5, or 10 years is rather
unlikely, or even infeasible.
Figure 1 (2) shows the cumulative stock market performance of value-weighted (equal-weighted)
portfolios for the two groups. Both figures document that firms in the High Sustainability group
significantly outperform firms in the Low Sustainability group. Investing $1 in the beginning of 1993 in a
value-weighted (equal-weighted) portfolio of High Sustainability firms would have grown to $22.6
($14.3) by the end of 2010. In contrast, investing $1 in the beginning of 1993 in a value-weighted (equalweighted) portfolio of control firms would have only grown to $15.4 ($11.7) by the end of 2010. Table 7
presents estimates from a four-factor model that controls for the market, size, book-to-market, and
momentum factors. We find that both portfolios exhibit statistically significant positive abnormal
performance relative to the market. However, we note that this might be because for both samples we
have chosen companies that survived and operated throughout the early 1990s and until the late 2000s.
The better performance of the firms in both samples compared to the rest of the market may be attributed,
to a considerable extent, to this survivorship bias. However, the relative performance difference between
the two groups is not affected by this bias since both groups are equally likely to have survived, by
construction of our sample. Accordingly, we find that the annual abnormal performance is higher for the
High Sustainability group compared to the Low Sustainability group by 4.8% (significant at less than 5%
level) on a value-weighted base and by 2.3% (significant at less than 10% level) on an equal-weighted
18
base.In fact, when we examine the performance of the two portfolios, we find that The High
Sustainability portfolio outperforms the control portfolio in 11 out of the 18 years. In addition, the High
Sustainability portfolio exhibits lower volatility. Whereas the standard deviation of monthly abnormal
returns is 1.43% and 1.72% on a value-weighted and equal-weighted base, respectively for the High
Sustainability group, the corresponding estimates for the Low Sustainability group are 1.72% and 1.79%.
To ensure that our results are not driven by long-run mean reversion in equity prices (Poterba and
Summers, 1988) or accounting profitability (Fama and French, 2000), we also examine the performance
of the two groups for the three years before 1993 (untabulated). We find that the two groups exhibit very
similar performance throughout these three years: cumulative stock returns are higher for the High
Sustainability group by only 1%. Similarly, cumulative ROA is higher for the Low Sustainability group
by only 0.04%, and cumulative ROE is higher for the High Sustainability group by 0.03%. This result is
consistent with our previous finding that matching in any one of the years between 1990 and1993 has
little impact on the composition of the pairs.
Overall, we find evidence that firms in the High Sustainability group are able to significantly
outperform their counterparts in the Low Sustainability group. This finding suggests that companies can
adopt environmentally and socially responsible policies without sacrificing shareholder wealth creation.
In fact, the opposite appears to be true: High Sustainability firms generate significantly higher stock
returns, suggesting that indeed the integration of such issues into a company’s business model and
strategy may be a source of competitive advantage for a company in the long-run. A more engaged
workforce, a more secure license to operate, a more loyal and satisfied customer base, better relationships
with stakeholders, greater transparency, a more collaborative community, and a better ability to innovate
may all be contributing factors to this potentially persistent superior performance in the long-term.
Sector Analysis
To shed some light on the underlying mechanisms that generate this outperformance, we construct a
cross-sectional model where the dependent variable is the alpha for each firm from the four-factor model
and the independent variable is an indicator variable for whether a firm is a member of the High
Sustainability group. We interact this variable with three additional indicator variables, each representing
sectors where we expect this outperformance to be more pronounced. The first moderator is an indicator
variable that takes the value of one for firms that are in business-to-consumer (B2C) sectors and zero for
firms that are in business-to-business (B2B) sectors. We expect that High Sustainability firms will
outperform their counterparts more in B2C businesses. In B2C businesses, individual consumers are the
customers, in contrast to B2B businesses where companies and governments are the customers. The
sensitivity of individual consumers to the company’s public perception is higher (Corey, 1991; Du,
Bhattacharya, and Sen 2007; Lev, Petrovits, and Radhakrishnan, 2010) and, as a result, the link between
19
sustainability and greater customer satisfaction, loyalty, and buying decisions should be stronger in B2C
businesses.
The second moderator is an indicator variable that takes the value of one for firms that are in
sectors where competition is predominantly driven by brand and reputation. Competing in such industries
usually requires employing high quality human capital for developing new products and sophisticated
marketing campaigns, and investment in continuous and rapid innovation. In these sectors, we expect that
the link between sustainability and attracting better employees, attaining higher levels of innovation, and
the management of reputational risk will be stronger. We proxy for sectors where brands and reputation
are relatively more important by constructing an indicator variable taking the value of one for sectors that
score at the fourth quartile of the market-to-book ratio in 1993 across all companies.
Finally, the third moderator is an indicator variable that takes the value of one for sectors where
firms’ products significantly depend upon extracting large amounts of natural resources (e.g. oil and gas,
chemicals, industrial metals, and mining). Particularly in recent years, firms in these sectors have been
subject to intense public scrutiny and many times have been in conflict with their local communities.
Moreover, environmental impact and resource scarcity are increasingly pressing social issues that have
increased regulation and put pressure on companies to minimize their environmental impact and become
more resource efficient. Therefore, we expect the link between sustainability and a more secure license to
operate, better community relations, and commercial benefits from a smaller environmental impact and
resource efficiency to be stronger in these sectors.
Table 8 presents the results from the cross-sectional model. In all specifications we include sector
fixed effects. In the first column, the model includes as an independent variable only the indicator
variable for High Sustainability firms. As expected, the coefficient is positive and significant. In the
second column we introduce the interaction terms with the moderators variables. All three coefficients on
the interaction terms are positive, as predicted. The coefficients on High Sustainability x B2C and High
Sustainability x Brand are significant at the 5% level. The coefficient on High Sustainability x Natural
Resources is significant at the 10% level. High Sustainability firms in B2C or Brand sectors outperform
their counterparts in 13 out of 18 calendar years whereas High Sustainability firms in the Natural
Resources sector outperform their counterparts in 11 out of 18 years. Overall, the results in Table 8
support our predictions that firms that integrated social and environmental issues in their business model
and strategy have benefited relatively more in B2C sectors and in sectors where companies compete on
the basis of brands and human capital, and where firms’ products depend on extracting large amounts of
natural resources.
Alternative Explanations
Alternative Explanation I: Price Pressure from SRI funds
20
We conclude this section by discussing alternative explanations. One potential explanation of higher stock
returns for High Sustainability firms is price pressure from the emergence of Socially Responsible
Investing (SRI). According to the Social Investment Forum, institutional investors that claim to
incorporate ESG data into their investment decisions had $162 billion in assets under management in
1995 and $2.5 trillion in 2010. However, the number of SRI funds that actually practice ESG integration
in a systematic way is lower with most SRI funds practicing ―negative screening‖ (i.e., excluding from
the investment universe specific sectors, such as tobacco), an investment strategy that does not affect our
results since the two groups have exactly the same industry composition. Nonetheless, to better
understand whether our results are driven by price pressure, and to mitigate concerns around the
inefficiency of stock prices as a performance metric, we examine the accounting performance of the two
groups of firms, which should not be affected by price pressure in stock markets. Moreover, the use of
accounting measures addresses concerns over stock price as a performance measure in the presence of
market inefficiencies that can prevent operating performance from being reflected in stock prices.
We find that High Sustainability firms outperform traditional ones when we consider accounting
rates of return. Figure 3 shows the cumulative performance of $1 of assets based on ROA. Investing $1 in
assets in the beginning of 1993 in a value-weighted (equal-weighted) portfolio of High Sustainability
firms would have grown to $7.1 ($3.5) by the end of 2010. In contrast, investing $1 in assets in a valueweighted (equal-weighted) portfolio of control firms would have grown to $4.4 ($3.3). Figure 4 shows the
cumulative performance of $1 of equity based on ROE. Investing $1 in book value of equity in the
beginning of 1993 in a value-weighted (equal-weighted) portfolio of High Sustainability firms would
have grown to $31.7 ($15.8) by the end of 2010. In contrast, investing $1 in book value of equity in a
value-weighted (equal-weighted) portfolio of control firms would have grown to $25.7 ($9.3).8 The
portfolio of High Sustainability firms outperforms the portfolio of control firms in 14 out of 18 years.
These results suggest that the stock market outperformance is founded on a solid outperformance in
accounting profitability.
Alternative Explanation II: Sustainability as a Luxury Good
Another alternative explanation is that the adoption of environmental and social policies is a luxury good
that firms can afford when they are performing well and therefore including in the Low Sustainability
group companies that throughout the years did not adopt these policies, is equivalent to selecting firms
that will underperform. To be more specific, a bias would only arise if an unidentified characteristic is
correlated with the sustainability policies, is uncorrelated with performance in the early 90s, and it is
8
It is worth noting that a substantial number of firms in the Low Sustainability group adopted a few environmental
and social policies throughout the 2000s. If this is not purely due to greenwashing, then this might bias our results
against finding performance differences across the two groups.
21
correlated with performance after 1993. However, the argument is inconsistent with the fact that in the
early 1990s the two sets of firms had statistically identical performance but had adopted very different
policies. It is also inconsistent with operating performance and leverage not being significant in the logit
model of propensity score matching, and with the fact that Low Sustainability firms have positive alphas
in the future. Moreover, when we test if past profitability is correlated with future adoption of policies
(changes in Sustainability Policies) we do not find a significant positive association. The coefficient on
past performance (e.g., three-year cumulative ROA, ROE, or stock returns) is slightly negative and
insignificant. Finally, the luxury good argument would predict that companies would drop these policies
in challenging times, such as during the financial crisis of 2008 and 2009. Contrary to this argument, we
find that in fact, companies slightly increased the number of policies during the financial crisis.
Sustainability Policies the equal-weighted policy index - increased from 0.28 in 2007 to 0.33, 0.34, and
0.36 in 2008, 2009, and 2010 respectively.
Alternative Explanation III: Omitted Risk Factor
The stock market outperformance documented here might be driven by an omitted risk factor that we
have been unable to identify and account for. Accordingly, we examine analyst surprises to annual
earnings announcements to differentiate between the omitted risk factor explanation and the market not
fully incorporating in stock prices the future profitability of High Sustainability firms. Table 9 shows the
results of analyzing forecast errors for the two groups of firms. We report results using as dependent
variables forecast errors (i.e. actual earnings minus the consensus forecasts) deflated by both the standard
deviation of analyst forecasts (SUE) and the absolute consensus forecast (%FE). Consistent with previous
research, we use the most recent consensus forecast error before the earnings announcement (Edmans,
2011). The coefficient on High Sustainability is positive and significant, suggesting that High
Sustainability firms have higher positive forecast errors and analysts being more positively surprised by
the future earnings of these firms.
Alternative Explanation IV: Survivorship Bias and Future Default Rates
In our primary sample selection we identified firms that had survived until the late 2000s because we
were interested in studying companies that have adopted environmental and social policies for multiple
consecutive years.. Because we have imposed the survivorship criterion for both groups of firms, it should
not affect the relative performance of the two groups. However, one remaining concern is that integration
of social and environmental issues might be a high risk-high return strategy that leads to a higher
probability of default and liquidation for a company. In untabulated results we calculate the index of
Sustainability Policies for all US firms with available data in 2003. Then we calculate the probability of
default and liquidation for each firm between 2004 and 2010. We do not observe any systematic relation
between Sustainability Policies and probability of default. Controlling for other determinants of default
22
we find that the coefficient on Sustainability Policies is negative and insignificant, suggesting that firms
that have adopted more environmental and social policies have a lower probability of default, although
this estimate is not reliably different than zero.
Alternative Explanation V: Corporate Governance as a Correlated Omitted Variable
Correlated omitted variables could be causing both the adoption of sustainability policies and future
performance. Our matching procedure attempted to create two statistically identical groups of firms, but
other characteristics that did not enter the matching algorithm could still be influencing the results. One
such variable could be corporate governance. Gompers et al. (2003) show that firms with more
shareholder-friendly governance provisions (G-index) outperformed their competitors in the 1990s. If
High Sustainability firms have a lower G-index, then the results documented here might be driven by
governance differences. However, we find that firms classified as High Sustainability have a higher Gindex (average is 9.6) compared to Low Sustainability firms (average is 8.2), suggesting that High
Sustainability firms have more powerful boards and less shareholder-friendly provisions. Moreover, we
analyzed the board characteristics of the two groups of firms in terms of independence and size. We did
not find any differences across the two groups.
8. Discussion
In this article we studied a matched sample of 180 companies, 90 of which we classified as High
Sustainability firms because they long ago adopted policies guiding their impact on the society and the
environment, while another 90 we classified as Low Sustainability firms because they had not. The Low
Sustainability firms largely correspond to the traditional model of corporate profit maximization in which
social and environmental issues are predominantly regarded as externalities. Often enough, responsibility
for forcing corporations to account for such externalities, whether positive or negative, rests with
governments and various laws and regulations that mandate certain kind of remedial actions. The High
Sustainability firms in contrast, not only pay attention to externalities but in fact, such firms are
characterized by distinct governance mechanisms which directly involve the board in sustainability issues
and link executive compensation to sustainability objectives; a much higher level of and deeper
stakeholder engagement, coupled with mechanisms for making it as effective as possible, including
reporting; a longer-term time horizon in their external communications which is matched by a larger
proportion of long-term investors; greater attention to nonfinancial measures regarding employees; a
greater emphasis on external environmental and social standards for selecting, monitoring and measuring
the performance of their suppliers; and a higher level of transparency in their disclosure of nonfinancial
information. In addition, during the 18-year period we studied, the High Sustainability firms
outperformed the Low Sustainability ones in terms of both stock market and accounting measures while
the market did not actually expect this outperformance. Our results also suggest that High Sustainability
23
firms benefited relatively more in B2C sectors and in sectors where companies compete on the basis of
brands and human capital, and where firms’ products depend on extracting large amounts of natural
resources.
We note that as with any quasi-experiment that lacks random assignment of treatment in a
laboratory setting, causality rather than correlation, between the independent variable and the dependent
variables of interest is up for debate. While we believe that our research design has many appealing
characteristics, we acknowledge the possibility that confounding factors might still exist. Future research
can examine the robustness and generalizability of our results to other settings, such as in other countries
or within the financial sector, and across different firm types, such as in private and smaller firms.
Moreover, an open question is whether our results generalize to firms that have adopted environmental
and social policies more recently. The difficulty in conducting such a study would be to distinguish
between companies that just mechanically mimic their peers with respect to sustainability policies and
those that adopt a more strategic approach.
More specifically, we suggest four areas for future research. The first area is to develop a better
understanding of the conditions under which companies decide to incorporate social and environmental
issues into their business model. The second area is the mechanisms by which such issues get integrated.
Two firms in the same conditions favoring the integration of social and environmental issues could differ
in the extent and speed with which they are able to do so. The third area for research is understanding
how the results presented in this paper vary across countries. In the presence of different legal, cultural,
and political institutions that affect corporate behavior with regards to sustainability (Ioannou and
Serafeim, 2010), one might expect High Sustainability firms to outperform even more than we have
documented here compared or to even underperform. Fourth, it would be useful to have deeper insights
into how differences in internal resource allocation resulting from the different characteristics of these
firms lead to the superior performance of the High Sustainability ones. For example, are these firms less
likely to cut back on R&D investments, lay off employees, and consolidate suppliers in economic down
cycles?
We note that, for the most part, even on those characteristics where the High Sustainability firms
were significantly different than their counterparts, the absolute percentages are relatively low. This
raises another interesting and important question, which is ―What is the optimal degree of adoption of
sustainability policies and practices?‖ Since sustainability involves tradeoffs, both across financial and
nonfinancial objectives, and between nonfinancial objectives themselves, such choices need to be well
understood in order to inform decision-making.
24
References
Aupperle, K. E., Carroll, A. B., and Hatfield, J. D. 1985. An empirical examination of the relationship between
corporate social responsibility and profitability. Academy of Management Journal, 28(2): 446-463.
Balotti, R. and Hanks, J. 1999. Giving at the office: a reappraisal of charitable contributions by corporations.
Business Lawyer 54 (3): 956-966.
Brochet, F., Loumioti, M., and Serafeim, G. 2012. Short-termism, Investor Clientele, and Firm Risk. Harvard
Business School Working Paper.
Brown, W., Helland, E., and Smith, K. 2006. Corporate philanthropic practices. Journal of Corporate Finance
12: 855-877.
Bushee, B.J. and Noe, C. 2001. Corporate Disclosure Practices, Institutional Investors, and Stock Return
Volatility. Journal of Accounting Research, 38: 171-202
Bushee, B. J. 2000. Do Institutional Investors Prefer Near-Term Earnings over Long-Run Value?
Contemporary Accounting Research, 18(2): 207-246
Cheng, B., Ioannou, I. and Serafeim, G. 2011. Corporate Social Responsibility and Access to Finance. HBS
Working Paper Series 11-130
Choi, J., and Wang, H., 2009. Stakeholder Relations and the Persistence of Corporate Financial Performance.
Strategic Management Journal, 30: 895-907
Clotfelter, C. 1985. Federal tax policy and charitable giving. Chicago: University of Chicago Press.
Cochran, P. L. and Wood, R. A. 1984. Corporate social responsibility and financial performance. The Academy
of Management Journal, 27(1): 42-56.
Du, S., Bhattacharya, C. B., and Sen, S. 2007. Reaping relational rewards from corporate social responsibility:
the role of competitive positioning. International Journal of Research in Marketing 24: 224-241.
Eccles, R. G., and Krzus, M. 2010. One Report: Integrated Reporting for a Sustainable Strategy. New York:
John Wiley and Sons, Inc.
Eccles, R. G., Edmondson, A. C., and Serafeim, G. 2011. A Note on Water. Harvard Business School Note
412-050.
Eccles, R., Krzus, M., and Serafeim, G. 2011 Market Interest in Nonfinancial Information. Harvard Business
School Working Paper, No. 12-018.
25
Edmans, A. 2011. Does the Stock Market Fully Value Intangibles? Employee Satisfaction and Equity Prices.
Journal of Financial Economics, 101: 621-640.
Fama, E., and French, K. 2000. Forecasting Profitability and Earnings. Journal of Business, 73 (2): 161-175.
Foo, L.M., 2007. Stakeholder engagement in emerging economies: considering the strategic benefits of
stakeholder engagement in a cross-cultural and geopolitical context. Corporate Governance, 7(4): 379-387.
Friedman, M., 1970. The social responsibility of business is to increase its profits. New York Times Magazine,
13 Sept., 32 (33): 122-126.
Freeman, R. E., 1984. Strategic Management: A stakeholder perspective. Boston, MA: Piman, United States.
Freeman, R. E., Harrison, J. S., and Wicks, A. C., 2007. Managing for stakeholders: survival, reputation, and
success. New Haven, CT: Yale University Press, United States.
Freeman, R. E., Harrison, J. S., Wicks, A. C., Parmar, B. L., de Colle, S., 2010. Stakeholder theory: the state
of the art. Cambridge University Press, New York, United States.
Galaskiewicz, J. 1997. An Urban Grants Economy Revisited: Corporate Charitable Contributions in the Twin
Cities, 1979-81, 1987-89. Administrative Science Quarterly, 42(3): 445-471.
Greening, D. W. and Turban, D. B., 2000. Corporate Social Performance as a Competitive Advantage in
Attracting a Quality Workforce. Business & Society, 39(3): 254-280.
Godfrey, P. C. 2005. The Relationship between Corporate Philanthropy and Shareholder Wealth: A Risk
Management Perspective. Academy of Management Review, 30(4): 777-798.
Gompers, P., Ishii, J., and Metrick, A. 2003. Corporate Governance and Equity Prices. Quarterly Journal of
Economics 118 (1): 107-156.
Govindarajan, V., and Gupta, A. 1985. Linking Control Systems to Business Unit Strategy: Impact on
Performance. Accounting, Organizations, and Society 10 (1): 51-66.
Healy, P., and Serafeim, G. 2011. Causes and Consequences of Firms’ Anticorruption Efforts. Harvard
Business School Working Paper.
Hill, C., and Jones, G. 2001. Strategic Management. Houghton Mifflin.
Hillman, A. J. and Keim, G. D., 2001. Shareholder Value, Stakeholder Management, and Social Issues: What's
the Bottom Line? Strategic Management Journal, 22(2): 125-139.
26
Ioannou, I. and Serafeim, G. 2010. What Drives Corporate Social Performance? The Role of Nation-level
Institutions. HBS Working Paper Series 11-016
Ioannou, I. and Serafeim, G. 2011. The Consequences of Mandatory Corporate Sustainability Reporting. HBS
Working Paper Series 11-100.
Ittner, C., Larcker, D., and Rajan, M. 1997. The Choice of Performance Measures in Annual Bonus Contracts.
The Accounting Review 72 (2): 231-255.
Jensen, M. and Meckling, W. 1976. Theory of the Firm: Managerial Behavior, Agency Costs and Ownership
Structure. Journal of Financial Economics 3(4): 305-360.
Jensen, M. 2001. Value Maximization, Stakeholder Theory and the Corporate Objective Function. Journal of
Applied Corporate Finance (fall): 8-21.
Jenter, D., Lewellen, K. and Warner, J.B. 2011.Security Issue Timing: What Do Managers Know, and When
Do They Know It? The Journal of Finance, 66(2):413-443.
Kaplan, R., and Norton, D. 1996. Using the Balanced Scorecard as a Strategic Management System. Harvard
Business Review 74, no. 1 (January-February ): 75-85.
Kaplan, R., and Norton D. 2008. Execution Premium: Linking Strategy to Operations for Competitive
Advantage. Harvard Business School Press.
Lev, B., Petrovits, C., and Radhakrishnan, S. Is doing good good for you? How corporate charitable
contributions enhance revenue growth. Strategic Management Journal 31: 182-200.
Jones, T.M., 1995. Instrumental stakeholder theory: A Synthesis of Ethics and Economics. Academy of
Management Review, 20: 404-437.
Laverty, K. J. 1996. Economic ―Short-Termism‖: The Debate, The Unresolved Issues, and The Implications
for Management Practice and Research. The Academy of Management Review, 21(3) 825-860.
Margolis, J. D. and Walsh, J. P. 2003. Misery loves companies: Rethinking social initiatives by business.
Administrative Science Quarterly, 48(2): 268-305.
Marquis, C., and Toffel, M. 2011. The Globalization of Corporate Environmental Disclosure: Accountability
or Greenwashing? Harvard Business School Working Paper, No. 11-115.
27
Margolis, J. D., Elfenbein, H. A., and Walsh, J. P. 2007. Does It Pay To Be Good? A Meta-Analysis and
Redirection of Research on the Relationship Between Corporate Social and Financial Performance, Working
paper.
McWilliams, A. and Siegel, D. 2001. Corporate social responsibility: A theory of the firm perspective.
Academy of Management Review, 26(1): 117-127.
Navarro, P. 1988. Why do corporations give to charity? Journal of Business, 61(1): 65-93.
Paine, Lynn Sharp. Value Shift: Why Companies Must Merge Social and Financial Imperatives to Achieve
Superior Performance. New York: McGraw-Hill, 2004.
Porter, M., and Kramer, M. 2011. Creating Shared Value. Harvard Business Review, 89 (1/2): 62-77.
Poterba, J.M. and Summers, L.H. 1988. Mean Reversion in Stock Prices: Evidence and Implications. Journal
of Financial Economics, 22(1): 27-59.
Sen, S., Gurhan-Canli, Z. And Morwitz, V. 2001. Withholding Consumption: A Social Dilemma Perspective
on Consumer Boycotts. Journal of Consumer Research, 28(3): 399-417.
Simnett, R., A. Vantraelen, and W. F. Chua. 2009. Assurance on Sustainability Reports: An International
Comparison. The Accounting Review 84 (May): 937–967.
Stein, J. 1989. Efficient Capital Markets, Inefficient Firms: A Model of Myopic Corporate Behavior
Quarterly Journal of Economics 104: 655-669.
Ullmann, A. A. 1985. Data in search of a theory: a critical examination of the relationships among social
performance, social disclosure, and economic performance of US firms. Academy of Management Review,
10(3): 540-557.
Waddock, S. A. and Graves, S. B. 1997. The corporate social performance-financial performance link.
Strategic Management Journal, 18(4): 303-319.
Wood, D. J. and Jones, R. E. 1995. Stakeholder mismatching: A theoretical problem in empirical research on
corporate social performance. International Journal of Organizational Analysis, 3: 229-267.
Zingales, L. 2000. In Search of New Foundations. The Journal of Finance, 55(4): 1623-1653.
28
Table 1 Summary statistics for two groups
Panel A: Sector composition of the sample
Sector
Oil & Gas Producers
Oil Equipment, Services & Distribution
Chemicals
Industrial Metals
Mining
Construction & Materials
Aerospace & Defense
General Industrials
Electronic & Electrical Equipment
Industrial Engineering
Industrial Transportation
Support Services
Automobiles & Parts
Beverages
Food Producers
Household Goods & Home Construction
Leisure Goods
Personal Goods
Health Care Equipment & Services
Pharmaceuticals & Biotechnology
Retailers
Media
Travel & Leisure
Fixed Line Telecommunications
Mobile Telecommunications
Electricity
Gas, Water & Multi-utilities
Software & Computer Services
Technology Hardware & Equipment
Total
29
% of sample
4.4%
3.3%
5.6%
1.1%
1.1%
1.1%
1.1%
4.4%
2.2%
3.3%
1.1%
1.1%
3.3%
1.1%
4.4%
3.3%
3.3%
2.2%
7.8%
4.4%
5.6%
3.3%
3.3%
2.2%
1.1%
6.7%
3.3%
5.6%
8.9%
100.0%
Panel B: Firm characteristics across two groups at the year of matching (1993)
Sustainability
Low
High
p-value diff
N
90
90
Total assets
Average
St. Dev.
8,182
28,213
8,591
22,230
0.914
ROA
Average
7.54
7.86
0.781
St. Dev.
8.02
7.54
ROE
Average
10.89
11.17
0.919
St. Dev.
20.61
16.15
Leverage
Average
0.57
0.56
0.726
St. Dev.
0.19
0.18
Turnover
Average
1.05
1.02
0.703
St. Dev.
0.62
0.57
MTB
Average
3.41
3.44
0.927
St. Dev.
2.18
1.88
Panel A: Frequency tabulation of sector membership for the 180 firms in the sample.
Panel B: All measures are calculated at the end of fiscal year in 1993. Averages and standard deviations across the High Sustainability and the Low Sustainability group are
presented. Each group includes 90 firms. ROA is net income plus net interest expense after tax over total assets. ROE is net income over shareholder’s equity. Leverage is total
liabilities over total assets. Turnover is sales over assets. MTB is stock price over book value of equity per share. p-value is derived from a test of the equality of the means across
the two groups.
30
Table 2
Governance
Panel A: Frequency Analysis of Governance
Governance
Board
Formal Board Responsibility / Corporate Citizenship
Sustainability committee
Compensation
Variable Compensation Metrics / Social Metrics
Variable Compensation Metrics / Environmental Metrics
Variable Compensation Metrics / External Perception Metrics
Sustainability
Low
High
Difference
p-value
21.6%
14.7%
52.7%
40.9%
<0.001
<0.001
21.6%
8.1%
10.8%
35.1%
17.6%
32.4%
0.022
0.011
0.004
This panel reports the frequency of stakeholder engagement practices across the two groups, low and high sustainability. p-value
is derived from a test of the equality of the frequencies across the two groups. ―Formal Board Responsibility / Sustainability‖ is
the percentage of companies that the board of directors explicitly assumes formal responsibility over corporate social
responsibility or sustainability. ―Sustainability committee‖ is the percentage of companies with a separate sustainability board
committee. ―Variable Compensation Metrics / Social Metrics‖ is the percentage of companies with pre-defined corporate social
indicators (e.g. corporate Health & Safety figure) relevant for the variable compensation of Executive / Top Management.
“Variable Compensation Metrics / Environmental Metrics” is the percentage of companies with pre-defined corporate
environmental indicators (e.g. corporate Emission reduction) relevant for the variable compensation of Executive / Top
Management. “Variable Compensation Metrics / External Perception Metrics” is the percentage of companies with pre-defined
corporate external perception indicators (e.g. reputational risks, customer satisfaction, feedback from stakeholder engagement)
relevant for the variable compensation of Executive / Top Management.
Panel B: Multivariate Analysis of Governance
Parameter
Intercept
High Sustainability
Size
MTB
ROA
Sector f.e.
N
Adj R-squared
Estimate
-0.773
0.144
0.084
-0.006
0.011
Yes
180
37.9%
p-value
0.003
0.006
0.002
0.563
0.055
This panel reports coefficient estimates and the statistical significance of these coefficients from an OLS model where the
dependent variable is the percentage of governance mechanisms in Panel A that a firm has adopted. ―High Sustainability‖ is an
indicator variable that takes the value of one if a firm is included in the High Sustainability group. ―Size‖ is the natural logarithm
of market capitalization at the end of 2009. ―MTB‖ is stock price over book value of equity per share at the end of 2009. ―ROA‖
is net income plus net interest expense after tax over total assets at the end of 2009. Standard errors are robust to
heteroscedasticity.
31
Table 3
Stakeholder Engagement
Panel A: Frequency Analysis of Stakeholder Engagement
Sustainability
Low
High
Stakeholder Engagement
Prior
Opportunities Risks Examination
Stakeholder Identification
Training
During
Concerns
Grievance Mechanism
Common Understanding
Scope Agreement
Targets
After
Board Feedback
Result Reporting
Public Reports
Difference
p-value
2.7%
10.8%
0.0%
31.1%
45.9%
14.9%
<0.001
<0.001
<0.001
2.7%
2.7%
13.5%
8.1%
0.0%
32.4%
18.9%
36.5%
36.5%
16.2%
<0.001
<0.001
<0.001
<0.001
<0.001
5.4%
0.0%
0.0%
32.4%
31.1%
20.3%
<0.001
<0.001
<0.001
This panel reports the frequency of stakeholder engagement practices across the two groups, low and high sustainability. p-value
is derived from a test of the equality of the frequencies across the two groups. ―Opportunities Risks Examination‖ is the
percentage of companies that undertake an examination of costs, opportunities and risks prior to a particular stakeholder
engagement. ―Stakeholder Identification‖ is the percentage of companies that identify issues and stakeholders that appear to be
most important for long-term success. ―Training‖ is the percentage of companies with local managers getting training in
stakeholder engagement. ―Concerns‖ is the percentage of companies for which the process of engagement ensures that all
stakeholders can rise their concerns. ―Grievance Mechanism‖ is the percentage of companies where grievance mechanisms are
agreed upon by all involved parties. ―Common Understanding‖ is the percentage of companies which formally pursue a
development of a common understanding of issues relevant to the underlying problem, such as technical terms. ―Scope
Agreement‖ is the percentage of companies that pursue mutual agreement on the type of engagement (type of meetings such as
group meetings, one-on-ones, frequency of meetings, exchange of information, roles of each party etc.) ―Targets‖ is the
percentage of companies with stakeholder engagement that targets are set and agreed upon by all involved parties. ―Board
Feedback‖ is the percentage of companies for which feedback from stakeholders is provided to board/supervisory board and/or
senior directors and/or compliance and/or communication department. ―Result Reporting‖ is the percentage of companies with
results of the engagement process being reported to the stakeholders involved. ―Public Reports‖ is the percentage of companies
that make the results of the engagement process publicly available.
32
Panel B: Multivariate Analysis of Stakeholder Engagement
Parameter
Intercept
High Sustainability
Size
MTB
ROA
Sector f.e.
N
Adj R-squared
Estimate
-0.428
0.245
0.073
-0.011
0.003
Yes
180
37.9%
p-value
0.167
<.0001
0.013
0.255
0.661
This panel reports coefficient estimates and the statistical significance of these coefficients from an OLS model where the
dependent variable is the percentage of stakeholder engagement mechanisms in Panel A that a firm has adopted. ―High
Sustainability‖ is an indicator variable that takes the value of one if a firm is included in the High Sustainability group. ―Size‖ is
the natural logarithm of market capitalization at the end of 2009. ―MTB‖ is stock price over book value of equity per share at the
end of 2009. ―ROA‖ is net income plus net interest expense after tax over total assets at the end of 2009. Standard errors are
robust to heteroscedasticity.
33
Table 4
Long-term Orientation
Panel A: Univariate Analysis of Long-term Orientation
Sustainability
Low
High
-5.31
-2.29
0.96
1.08
Measures of long-term orientation
Long-term vs. Short-term Investors
Long-term vs. Short-term Discussion
Difference
p-value
<0.001
0.030
This table reports the average long-term orientation of the two groups, low and high sustainability. p-value is derived from a test
of the equality of the means across the two groups. ―Long-term vs. Short-term Investors‖ is the percentage of shares outstanding
held by dedicated investors minus the percentage of shares held by transient investors. This investor classification is based on the
one used in Bushee (2001) and Bushee and Noe (2000). ―Long-term vs. Short-term Discussion‖ is the ratio of long-term over
short-term keywords included in transcripts of discussions between the management and sell-side analysts in conference calls.
Panel B: Multivariate Analysis of Long-term Orientation
Dependent variable
Parameter
Intercept
High Sustainability
Size
MTB
ROA
Sector f.e.
Year f.e.
N
Adj R-squared
Long-term vs. Shortterm Investors
Estimate
p-value
-39.348
<.0001
3.012
0.004
2.158
<.0001
0.014
0.810
2.277
0.740
Yes
Yes
830
8.8%
Long-term vs. Shortterm Discussion
Estimate
p-value
1.897
0.000
0.038
0.070
-0.053
0.093
0.004
0.017
-0.018
0.971
Yes
Yes
980
10.4%
This panel reports coefficient estimates and the statistical significance of these coefficients from an OLS model where the
dependent variable is Long-term vs. Short-term Investors or Long-term vs. Short-term Discussion in Panel A. ―High
Sustainability‖ is an indicator variable that takes the value of one if a firm is included in the High Sustainability group. ―Size‖ is
the natural logarithm of market capitalization at fiscal year-end. ―MTB‖ is stock price over book value of equity per share at fiscal
year-end. ―ROA‖ is net income plus net interest expense after tax over total assets at fiscal year-end. The data cover years 20022008. Standard errors are robust to heteroscedasticity and clustered at the firm level.
34
Table 5
Measurement of Nonfinancial Information
Panel A: Employees
Sustainability
Low
High
16.2%
54.1%
26.3%
77.4%
26.3%
64.5%
89.5%
95.2%
Employees
HR Performance Indicators / Nonfinancial
KPI Labor / EHS Fatalities Tracking
KPI Labor / EHS Near Miss Tracking
KPI Labor / EHS Performance Tracking
Difference
p-value
<0.001
<0.001
<0.001
0.871
This table reports the frequency of stakeholder engagement practices across the two groups, low and high sustainability. p-value
is derived from a test of the equality of the frequencies across the two groups. ―HR Performance Indicators/Nonfinancial‖ is the
percentage of companies that use HR Performance Indicators / Nonfinancial (e.g. number of hours spent in trainings, companyspecific skills categorization) to measure execution of skill mapping and development strategy. ―KPI Labor/EHS Fatalities
Tracking‖ is the percentage of companies that use fatalities tracking to follow labor relations issues. ―KPI Labor/EHS Near Miss
Tracking‖ is the percentage of companies that use near miss tracking to follow labor relations issues. ―KPI Labor/EHS
Performance Tracking‖ is the percentage of companies that use health and safety performance tracking to follow labor relations
issues.
Panel B: Customers
Customers
Customer Lifestyle
Geographical Segmentation
Potential Lifetime Value
Customer Generated Revenues
Historical Sales Trends
Products Bought
Cost Of Service
Sustainability
Low
High
2.7%
5.4%
10.8%
18.9%
2.7%
8.1%
8.1%
18.9%
8.1%
16.2%
8.1%
14.9%
2.7%
6.8%
Difference
p-value
0.461
0.101
0.164
0.041
0.100
0.194
0.279
This table reports the frequency of stakeholder engagement practices across the two groups, low and high sustainability. p-value
is derived from a test of the equality of the frequencies across the two groups. ―Customer Lifestyle‖ is the percentage of
companies that use customer lifestyle to segment customers in the company’s CRM database. ―Geographic Segmentation‖ is the
percentage of companies that use geographic segmentation to segment customers in the company’s CRM database. ―Potential
Lifetime Value‖ is the percentage of companies that use the potential lifetime value to business to segment customers in the
company’s CRM database. ―Customer Generated Revenues‖ is the percentage of companies that use the revenues generated by
customers to segment customers in the company’s CRM database. ―Historical Sales Trends‖ is the percentage of companies that
use historical sales trends to segment customers in the company’s CRM database. ―Products Bought‖ is the percentage of
companies that use products/services bought to segment customers in the company’s CRM database. ―Cost of Service‖ is the
percentage of companies that use the costs of services to segment customers in the company’s CRM database.
35
Panel C: Suppliers
Sustainability
Low
High
Suppliers
Environmental
EMS
Environmental Production Standards
Environmental Data Availability
Environmental Policy
Product LCA
Social
Human Right Standards
OHS Standards
Grievance Process
Labor Standards
Standards
International Standards Compliance
National Standards Compliance
Difference
p-value
18.2%
25.7%
0.0%
0.0%
0.0%
50.0%
45.6%
12.3%
17.4%
6.6%
<0.001
<0.001
0.018
<0.001
0.052
5.7%
25.7%
0.0%
8.1%
17.4%
62.9%
8.1%
18.6%
<0.001
<0.001
0.039
0.020
0.0%
8.1%
12.3%
14.9%
<0.001
0.057
This table reports the frequency of stakeholder engagement practices across the two groups, low and high sustainability. p-value
is derived from a test of the equality of the frequencies across the two groups. ―EMS‖ is the percentage of companies that use
EMS in their certification/audit/verification process. ―Environmental Production Standards‖ is the percentage of companies that
use environmental production standards to select and evaluate the company’s key suppliers and services providers.
―Environmental Data Availability‖ is the percentage of companies that use environmental data availability to select and evaluate
the company’s key suppliers and services providers. ―Environmental Policy‖ is the percentage of companies that use
environmental policy to select and evaluate the company’s key suppliers and services providers. ―Product LCA‖ is the percentage
of companies that use product lifecycle impact assessment to select and evaluate the company’s key suppliers and services
providers. ―Human Rights Standards‖ is the percentage of companies that use human rights standards (such as forced, slave
labor, child labor) to select and evaluate the company’s key suppliers and services providers. ―OHS Standards‖ is the percentage
of companies that use occupational health & safety to select and evaluate the company’s key suppliers and services providers.
―Grievance Process‖ is the percentage of companies that use grievance process implementation to select and evaluate the
company’s key suppliers and services providers. ―Labor Standards‖ is the percentage of companies that use labor
standards/requirements to select and evaluate the company’s key suppliers and services providers. ―International Standards
Compliance‖ is the percentage of companies that use international standards compliance to select and evaluate the company’s
key suppliers and services providers. ―National Standards Compliance‖ is the percentage of companies that use national
standards compliance to select and evaluate the company’s key suppliers and services providers.
36
Panel D: Assurance
Sustainability
Low
High
1.4%
11.1%
Assurance
Sustainability report external audit
Assurance Provision Process
Information Collection Review
Data Aggregation Review
Document Review
Relevant Management Interviews
Mapping against Standards
Auditor Competency Disclosure
Relevant Management Discussions
Sample Site Visits
Stakeholder Consultation
Distribution Network Quality
External Audits
Standardized External Audits
Internal Audits
Difference
p-value
0.017
5.4%
5.4%
5.4%
5.4%
2.7%
2.7%
5.4%
2.7%
0.0%
14.9%
14.9%
14.9%
12.2%
16.2%
5.4%
14.9%
12.2%
5.4%
0.058
0.058
0.058
0.089
0.031
0.589
0.058
0.058
0.131
8.1%
5.4%
5.4%
12.2%
12.2%
13.5%
0.221
0.058
0.046
This table reports the frequency of stakeholder engagement practices across the two groups, low and high sustainability. p-value
is derived from a test of the equality of the frequencies across the two groups. ―Sustainability report external audit‖ is the
percentage of companies with a public sustainability report that is assured by a third party. ―Information Collection Review‖ is
the percentage of companies that use a review of internal processes of information generation and collection as an element in the
company’s assurance provision process. ―Data Aggregation Review‖ is the percentage of companies that use a review of data
aggregation processes as an element in the company’s assurance provision process. ―Document Review‖ is the percentage of
companies that use a review of documents as an element in the company’s assurance provision process. ―Relevant Management
Interviews‖ is the percentage of companies that conduct interviews with management responsible for the information gather
process at the corporate level as an element in the company’s assurance provision process. ―Mapping Against Standards‖ is the
percentage of companies that map against relevant external standards and programs, including AA1000 and the Global Reporting
Initiative, as an element in the company’s assurance provision process. ―Auditor Competency Disclosure‖ is the percentage of
companies that use disclosure of competencies of assurance providers as an element in the company’s assurance provision
process. ―Relevant Management Discussions‖ is the percentage of companies that incorporate relevant management discussions
as an element in the company’s assurance provision process. ―Sample Site Visits‖ is the percentage of companies that conduct use
sample site visits as an element in the company’s assurance provision process. ―Stakeholder Consultation‖ is the percentage of
companies that consultations with stakeholders as an element in the company’s assurance provision process. ―External Audits‖ is
the percentage of companies that use external audits as an element in the company’s assurance provision process. ―Internal
Audits‖ is the percentage of companies that use internal audits as an element in the company’s assurance provision process.
37
Panel E: Multivariate Analysis of Nonfinancial Measurement
Parameter
Intercept
High Sustainability
Size
MTB
ROA
Sector f.e.
N
Adj R-squared
Employees
Estimate
p-value
-0.206
0.394
0.174
0.000
0.045
0.040
0.008
0.331
-0.005
0.257
Yes
180
30.3%
Customers
Estimate p-value
0.222
0.154
0.013
0.681
0.014
0.388
0.002
0.843
-0.008
0.046
Yes
180
41.9%
Suppliers
Estimate
p-value
-0.266
0.060
0.089
0.001
0.041
0.002
0.007
0.176
-0.003
0.343
Yes
180
45.1%
Assurance
Estimate
p-value
0.272
0.329
0.115
0.020
0.035
0.122
-0.001
0.887
-0.016
0.007
Yes
180
49.0%
This panel reports coefficient estimates and the statistical significance of these coefficients from an OLS model where the
dependent variable is the percentage of employee, customer, supplier, and assurance procedure in Panels A-D that each firm
adopted. ―High Sustainability‖ is an indicator variable that takes the value of one if a firm is included in the High Sustainability
group. ―Size‖ is the natural logarithm of market capitalization at the end of 2009. ―MTB‖ is stock price over book value of equity
per share at the end of 2009. ―ROA‖ is net income plus net interest expense after tax over total assets at the end of 2009. Standard
errors are robust to heteroscedasticity.
38
Table 6
Disclosure of Nonfinancial Information
Panel A: Univariate Analysis of Nonfinancial Disclosure
Nonfinancial disclosure
Quantity
ESG Disclosure - Bloomberg
ESG Disclosure - Thomson Reuters
Coverage
Sustainability report covers global activities
Integration
Nonfinancial vs. Financial Discussion
Social Data Integrated in Financial Reports
Environmental Data Integrated in Financial Reports
Sustainability
Low
Difference
High
p-value
17.86
36.91
29.90
46.38
<0.001
<0.001
8.3%
41.4%
<0.001
0.68
5.4%
10.8%
0.96
25.7%
32.4%
<0.001
0.008
0.011
―ESG Disclosure – Bloomberg‖ is the average disclosure score across the two groups, low and high sustainability. Bloomberg
calculates this score based on the percentage of sustainability datapoints each company discloses. The measure ranges from 0 to
100. ―ESG Disclosure – Thomson Reuters‖ is the average disclosure score across the two groups, low and high sustainability. We
calculated this score based on the percentage of sustainability datapoints each company discloses, using Thomson Reuters
ASSET4 data. The measure ranges from 0 to 100. ―Sustainability report covers global activities‖ is the percentage of companies
in each group that publishes a sustainability report that covers the global operations of the firm. ―Nonfinancial vs. Financial
Discussion‖ is the ratio of nonfinancial over financial keywords included in transcripts of discussions between the management
and sell-side analysts in conference calls. The higher this number the more emphasis on nonfinancial topics a management
places. ―Social Data Integrated in Financial Reports‖ is the percentage of companies in each group that integrated social KPIs
and narrative information in their financial reporting. ―Environmental Data Integrated in Financial Reports‖ is the percentage of
companies in each group that integrated environmental KPIs and narrative information in their financial reporting.
39
Panel B: Multivariate Analysis of Nonfinancial Disclosure
Dependent
variable
Parameter
Intercept
High Sustainability
Size
MTB
ROA
Sector f.e.
Year f.e.
N
Adj R-squared
ESG Disclosure Bloomberg
Estimate
-49.201
8.618
4.502
0.222
-45.407
Yes
Yes
464
32.1%
p-value
<.0001
<.0001
<.0001
0.095
<.0001
ESG Disclosure Thomson Reuters
Estimate
-38.002
9.759
2.236
0.238
-12.070
Yes
Yes
1314
30.8%
p-value
<.0001
<.0001
<.0001
0.095
<.0001
Nonfinancial
vs. Financial
Discussion
Estimate p-value
-2.864 <.0001
0.180
0.043
0.231 <.0001
0.000
0.841
-0.597
0.243
Yes
Yes
980
11.0%
Sustainability report
covers global activities
Social Data Integrated
in Financial Reports
Odds ratio
Odds ratio
4.279
2.158
0.972
0.004
Yes
Yes
2072
22.9%
p-value
<0.001
<0.001
0.069
<0.001
5.395
1.325
1.030
1.014
Yes
No
180
13.2%
p-value
0.031
0.238
0.744
0.722
Environmental Data
Integrated in Financial
Reports
Odds ratio
p-value
4.143
1.270
1.160
0.969
Yes
No
180
12.5%
0.028
0.275
0.070
0.461
This panel reports coefficient estimates and the statistical significance of these coefficients from OLS and logistic models where the dependent variable is one of the variables
reported in Panel A. Data for years 2005-2010 are used for the ―ESG Disclosure – Bloomberg‖ analysis. Data for years 2002-2010 are used for the ―ESG Disclosure - Thomson
Reuters‖ and ―Sustainability report covers global activities‖ analyses. Data for years 2002-2008 are used for the ―Nonfinancial vs. Financial Discussion‖ analysis. For the last two
analyses data from year 2009 are used. ―High Sustainability‖ is an indicator variable that takes the value of one if a firm is included in the High Sustainability group. ―Size‖ is the
natural logarithm of market capitalization at the end of 2009. ―MTB‖ is stock price over book value of equity per share at the end of 2009. ―ROA‖ is net income plus net interest
expense after tax over total assets at the end of 2009. Standard errors are robust to heteroscedasticity and clustered at the firm level (first four columns).
40
Table 7
Stock Market Performance
Value-weighted
Equal-weighted
Sustainability
Low
High
Low
High
Parameter
Estimate
p-value
Estimate
p-value
Estimate
p-value
Estimate
p-value
Intercept
MKTRF
SMB
HML
UMD
N
Adj R-squared
0.0059
0.9839
-0.2076
0.1982
-0.0156
216
85.6%
<.0001
<.0001
<.0001
0.001
0.642
216
0.0096
0.9360
-0.1776
-0.2727
-0.0266
216
86.6%
<.0001
<.0001
0.002
<.0001
0.427
216
0.0039
0.9977
0.1598
0.4053
-0.1436
216
88.9%
0.004
<.0001
0.001
<.0001
<.0001
216
0.0057
0.9557
0.0366
0.2204
-0.1239
216
91.0%
<.0001
<.0001
0.367
<.0001
<.0001
216
This table provides the estimates from a Fama-French four-factor model augmented by the Carhart momentum factor. The dependent variable is the monthly portfolio stock return
for low or high sustainability minus the risk-free rate for that month. MKTRF is the value-weighted market return minus the risk-free rate for that month. SMB is the monthly
return on a hedge portfolio that mimics the stock returns of small minus large firms. HML is the monthly return on a hedge portfolio that mimics the stock returns of low MTB
minus high MTB firms. UMD is the monthly return on a hedge portfolio that mimics the stock returns of high prior returns minus low prior returns firms. The intercept represents
the abnormal stock return for the average month. We estimate the model for the period 1993-2010.
41
Table 8
Stock Market Performance and Sector Membership
Parameter
Estimate
Intercept
High Sustainability
High Sustainability x B2C
High Sustainability x Brand
High Sustainability x Natural Resources
Sector fixed effects
Adj R-squared
N
0.0118
0.0019
Yes
32.5%
180
p-value
<.0001
0.014
Estimate
0.0124
-0.0008
0.0040
0.0038
0.0018
Yes
35.9%
180
p-value
<.0001
0.584
0.031
0.044
0.100
This table presents estimates from OLS models where the dependent variable is the alpha for each firm from a Fama-French four-factor model augmented by the Carhart
momentum factor. ―High Sustainability‖ is an indicator variable that takes the value of one if a firm is included in the High Sustainability group. ―B2C‖ is an indicator variable that
takes the value of one for firms that operate in sectors where the customers are individual people instead of companies or governments. ―Brand‖ is an indicator variable that takes
the value of one for firms that operate in sectors that rank at the fourth quartile of market-to-book ratios in 1993. ―Natural resources‖ is an indicator variable that takes the value of
one for firms that operate in sectors that require the extraction of large amounts of natural resources. Standard errors are robust to heteroscedasticity.
42
Table 9
Forecast Errors
SUE
Parameter
Estimate
%FE
p-value
Estimate
p-value
Intercept
0.860
0.008
0.082
0.000
High Sustainability
0.245
0.007
0.008
0.068
Size
-0.022
0.526
-0.006
0.006
MTB
0.899
0.003
0.011
0.429
Sector f.e.
Yes
Yes
Year f.e.
Yes
Yes
N
2406
2680
Adj R-squared
2.8%
3.7%
This panel reports coefficient estimates and the statistical significance of these coefficients from OLS models where the dependent variable is the difference between actual
earnings and the mean estimate over the standard deviation of the estimates (SUE) or the difference between actual earnings and the mean estimate over the mean estimate (%FE).
―High Sustainability‖ is an indicator variable that takes the value of one if a firm is included in the High Sustainability group. ―Size‖ is the natural logarithm of market
capitalization at fiscal year-end. ―MTB‖ is stock price over book value of equity per share at fiscal year-end. Standard errors are robust to heteroscedasticity and clustered at the
firm level.
43
Figure 1
Evolution of $1 invested in the stock market in value-weighted portfolios
25.00
20.00
15.00
Low
10.00
High
5.00
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
0.00
Figure 2
Evolution of $1 invested in the stock market in equal-weighted portfolios
15.00
12.00
9.00
Low
High
6.00
3.00
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
0.00
44
Figure 3
Evolution of $1 of assets based on Return-on-Assets
8
7
6
5
Low VW
4
High VW
3
Low EW
2
High EW
1
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
0
Figure 4
Evolution of $1 of equity based on Return-on-Equity
35
30
25
Low VW
20
High VW
15
Low EW
High EW
10
5
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
0
45
Appendix
Name
Description
Bonus Plan for Employees/Employees
Community/Policy I
Community/Policy II
Diversity and Opportunity/Policy
Emission Reduction Policy Elements/Emissions
Emission Reduction/CO2 Reduction
Emission Reduction/Transportation Impact Reduction
Employee welfare
Employment Quality/Policy I
Employment Quality/Policy II
Environmental Supply Chain Management
Generous Fringe Benefits
Health & Safety /Policy
Human Rights Contractor
Human Rights/Policy I
Human Rights/Policy II
Internal Promotion
Management Training
Positive Discrimination
Product Impact Minimization
Product Innovation
Product Responsibility/Policy I
Product Responsibility/Policy II
Resource Efficiency/Energy Efficiency Policy
Resource Efficiency/Water Efficiency Policy
Training and Development/Policy
Waste Reduction Total
Does the company provide a bonus plan to most employees?
Does the company have a policy to strive to be a good corporate citizen or endorse the Global Sullivan Principles?
Does the company have a policy to respect business ethics or has the company signed the UN Global Compact or follow the OECD guidelines?
Does the company have a diversity and equal opportunity policy?
Does the company have a policy to reduce emissions?
Does the company shows an initiative to reduce, reuse, recycle, substitute, phased out or compensate CO2 equivalents in the production process?
Does the company have initiatives to reduce the environmental impact of transportation of its products or its staff?
Does the company have a work-life balance policy?
Does the company have a competitive employee benefits policy or ensuring good employee relations within its supply chain?
Does the company have a policy for maintaining long term employment growth and stability?
Does the company use environmental criteria (ISO 14000, energy consumption, etc.) in the selection process of its suppliers or sourcing partners?
Does the company claim to provide its employees with a pension fund, health care or other insurances?
Does the company have a policy to improve employee health & safety within the company and its supply chain?
Does the company show to use human rights criteria in the selection or monitoring process of its suppliers or sourcing partners?
Does the company have a policy to guarantee the freedom of association universally applied independent of local laws?
Does the company have a policy for the exclusion of child, forced or compulsory labor?
Does the company claim to favor promotion from within?
Does the company claim to provide regular staff and business management training for its managers?
Does the company promote positive discrimination?
Does the company design product features and applications/services that promote responsible, efficient, cost-effective and environmentally preferable use?
Does the company have take-back procedures and recycling programs to reduce the potential risks of products entering the environment?
Does the company have a policy to protect customer health & safety?
Does the company have a products and services quality policy?
Does the company have a policy to improve its energy efficiency?
Does the company have a policy to improve its water efficiency?
Does the company have a policy to support the skills training or career development of its employees?
Does the company have initiatives to recycle, reduce, reuse, substitute, treat or phase out total waste?
This Appendix includes all environmental and social policies that are used to construct the Sustainability Policies index.
46